FRAND Rate-Setting: Convergence on Principles, Divergence in Practice

1. Introduction

Courts in different jurisdictions have determined fair, reasonable, and non-discriminatory (FRAND) rates for SEP licenses for almost 15 years, and the methodologies used in such determinations are increasingly well understood. Nevertheless, recent decisions illustrate how widely different courts’ determinations of FRAND rates for essentially the same license can still vary. In Samsung v. ZTE, the UK High Court determined a royalty of USD 392 million for the renewal of a prior ZTE-Samsung cross-license.¹

In contrast, the Chongqing Intermediate People’s Court reached a materially different outcome, holding that ZTE’s offer of USD 731 million for a six-year cross-license from 2024 to 2029 was FRAND. Even adjusted to the five-year term considered by the UK High Court, that figure would still exceed USD 600 million.² The Munich I Regional Court also took a different view, finding that ZTE could have demanded up to USD 798.6 million from Samsung while proposing USD 640 million as a settlement figure for a five-year cross-license.³ Lastly, the Mannheim Local Division from the Unified Patent Court (UPC) proposed settlement terms of USD 640 million for a cross-license until the end of 2028 and USD 730 million for a cross-license until the end of 2029.⁴ This divergence illustrates a central challenge of FRAND rate determinations: Even where courts agree on the general principles and the basic meaning of FRAND, outcomes depend significantly on specific methodological aspects, such as, among others, the selection of comparable licenses, the adjustments applied to account for differences between licenses, and the so-called “non-FRAND” factors.

Judicial decisions are the most important source for understanding how FRAND works in practice. Yet this growing body of caselaw is dispersed across jurisdictions, languages, legal traditions, and procedural settings. This makes it difficult to compare how courts have applied different methodologies across decisions, how they have dealt with evidentiary constraints, and how robust the underlying indicators are. Consequently, existing reviews in the literature have often focused on selected landmark decisions rather than a consistently assembled set of cases.

The European Patent Office (EPO), together with our team at the economic consultancy BRELA, has produced the most extensive methodological review of judicial FRAND determinations to date (click here to visit the new EPO & BRELA study). The study covers 65 court documents across seven jurisdictions and distinguishes between three categories: FRAND rate determinations, in which courts set a FRAND rate; FRAND rate assessments, in which courts decide whether a given rate or offer is FRAND or falls within a FRAND range; and decisions on the admissibility or suitability of FRAND determination methods, including U.S. Daubert rulings and selected court orders or guidelines from other jurisdictions.

“The study assembles the most comprehensive global corpus of judicial FRAND determinations to date, spanning 65 court decisions across seven jurisdictions.”⁵

The study systematically identifies the methodologies that courts use, the principles and frameworks within which they operate, and the points of convergence and divergence across jurisdictions. It also examines how courts apply these methodologies in practice and how they use patent data for apportionment, scaling, and the unpacking of comparable licenses, and how they assess essentiality, validity, and patent value. Importantly, the study does not seek to determine the legal status of different principles, methods, or arguments in any individual jurisdiction, but provides an empirical and descriptive account of the methodologies that courts have adopted and the challenges associated with different approaches.

This study complements a recent WIPO publication on FRAND valuation methodology (click here to visit the WIPO study).⁶ While the WIPO publication explains the economic rationale and properties of structured valuation methods, the study places greater emphasis on extant caselaw. It systematically reviews how courts in seven jurisdictions have used these methods in practice and highlights the methodological choices, challenges, and open questions that arise when economic valuation tools are applied in litigation.

2. Convergence in FRAND Methodology: Shared Principles, Frameworks, and Overall Methodologies

The review of the FRAND caselaw shows that courts have developed a more settled understanding of FRAND as a principle. Irrespective of whether courts apply contract or competition law, they generally frame FRAND around the same balance, namely, that SEP owners should receive fair compensation, while implementers should be given access to the standard. While earlier decisions emphasized the potential risks of royalty stacking and patent hold-up, the more recent caselaw identifies both hold-up and hold-out as potential concerns.

“Despite operating under different legal frameworks, courts across jurisdictions have converged on a common purpose for FRAND and a shared conceptual framework, and the focus has shifted from defining FRAND to applying it.”⁷

These widely shared principles are applied through theoretical frameworks that are themselves broadly shared. Courts in the United States have generally used a hypothetical-negotiation framework to adapt traditional patent damages analysis to the FRAND context, particularly through the modified Georgia-Pacific approach (see, e.g., Microsoft v. Motorola and In re Innovatio). UK courts have applied a similar principle through the willing licensor/willing licensee framework. Courts in the United Kingdom, the United States, and Germany have also aligned in rejecting an interpretation of the non-discrimination obligation that would automatically require every licensee to receive the same terms as the most favorable prior license.

As a result of this convergence on certain principles and frameworks, the focus of the caselaw has gradually moved away from abstract debates about FRAND principles, which dominated between 2013 and 2015, toward a more practical focus on methodology.

“Two methodologies dominate FRAND rate determinations – comparable licences and the top-down approach – with comparable licences emerging as the primary method in most cases and top-down used mainly as a cross-check.”⁸

There is further strong convergence around two approaches to FRAND determinations: comparable licenses and, to a lesser extent, top-down analysis, which courts have typically used as a cross-check. Courts most often rely on comparable licenses because existing agreements may be best-suited to show how market participants have priced access to relevant patented technologies. Top-down analysis plays a smaller but still important role, especially where courts lack reliable comparables (see, e.g., In re Innovatio in the U.S.; Samsung v. Apple in Japan; Siemens v. Xiaomi and Huawei v. Conversant in China, or the Frankfurt regional court decision in Samsung v ZTE in Germany) or use it to assess whether a rate derived from comparables appears plausible (e.g., Unwired Planet v. Huawei, or the Appeal Court decision in Optis v. Apple). Courts’ assessments of whether proposed licensing rates are FRAND show a pattern similar to rate determinations: courts primarily rely on comparable licenses, while top-down analysis is the second most commonly used method.

3. Challenges at the Operational Level: Selection and Unpacking of Comparable Licenses, Use of Patent Data

At the operational level, however, important questions remain contested. This is particularly visible in comparable license analysis. Courts in different jurisdictions follow the same basic steps, but each step entails numerous specific methodological choices that may lead to divergent outcomes.

The basic steps of a comparable license analysis are selection, unpacking, and adjustments. Firstly, courts must decide which agreements qualify as sufficiently comparable, taking into account the licensed technology, the parties’ identities and characteristics, the scope of the license, the timing of the agreement, and the commercial context of the negotiation. Secondly, they must also unpack the numerous value-relevant aspects of licensing agreements to derive an effective rate that can be compared across different licenses. Finally, courts must adjust for observed differences between the comparable license and the license at issue (including portfolio strength, geographic coverage, product scope, timing), and account for and correct possible “non-FRAND factors” (such as asymmetric bargaining power between the parties). The fact that the focus of the FRAND debate has now shifted to such detailed questions is, in itself, a sign of methodological progress, even though it also highlights that converging interpretations of FRAND’s meaning alone do not significantly reduce the room for parties’ disagreement over FRAND rates.

The study provides a detailed account of courts’ treatment of the different steps of the comparable license analysis. In particular, it reviews the proposed comparables from various FRAND cases and identifies and discusses the selection factors that parties and courts used to select the licenses to be relied upon. It also provides a detailed review of courts’ approaches to the unpacking of lump-sum payments, cross-licenses, releases for past infringement, and multi-technology licenses.

Another important focus of the study is the use of patent data in FRAND rate determinations and courts’ assessments of licensing offers. The focus on patent data is timely, as the EPO is creating new datasets and resources, which may address some of the numerous known limitations of existing patent data. The study’s focus on patent data is also in keeping with BRELA’s core expertise in this area, as reflected, e.g., in the role that Dr. Baron’s expert witness testimony has played in the UK High Court proceeding in Samsung v. ZTE.

The study shows that the use of patent data in FRAND determinations is not limited to any specific methodology. Patent data is proposed or used, for instance, for the apportionment step of a top-down approach, the selection and adjustment of comparable licenses (when the portfolio of patents that is licensed at issue is not the same as in the comparable license), the unpacking of comparable licenses into two unilateral rates, or adjustments of royalty rates to reflect evolutions in patent portfolio strength over time. Nevertheless, there are important and well-understood limitations of existing uses of patent data – and these limitations are shared between the different uses of patent data within different FRAND rate determination methodologies.

“There are many use cases for patent data in different FRAND rate determination methodologies, and they share many of the same methodological challenges.”

The study reviews different courts’ attempts to account for uncertainty regarding essentiality, validity, and value of different declared SEPs, and compares different approaches to patent families, pending applications, and expired patents. It also documents courts’ recognition of the limitation of simplistic counting methods; as many courts have recognized that the issue of portfolio evaluation “is not simply a numeric equation but the compensation must, within reasonable bounds, reflect the contribution.”⁹ Of course, measuring the contribution of patented inventions to the standard is very distinct from counting companies’ contributions to standards meetings – a method that shares many of the limitations of patent counting, and introduces additional problems (such as the fact that many such contributions are untethered to any patented inventions).

Overall, extant caselaw points to broad convergence on the main methodological approaches to FRAND determinations, but divergence and significant challenges remain at the operational level. Courts increasingly agree on the purpose of FRAND and on the main tools used to assess it, especially comparable licenses and top-down analysis. Today’s disputes increasingly concern the specific components of these methodologies, including the selection of comparables and adjustments to the observed unpacked rates, and the practical challenges of using patent data (such as variations in essentiality, validity, and value of different patents).

4. Divergence in FRAND Review: FRAND Rate Determinations versus FRAND Assessments

One of the main findings of the study is thus that the focus of FRAND disputes has moved from fundamental disagreements about the meaning of FRAND to the practical implementation. The direction in which the caselaw has evolved certainly is a testament to the advantages of the prevailing case-by-case approach, through which the substance of FRAND obligations is gradually clarified. Nevertheless, as illustrated by the divergence between outcomes in parallel proceedings in the same dispute, additional progress towards consensus at the most operational level is needed.

In parallel to this methodological convergence, different jurisdictions continue to differ in the type of FRAND review they undertake. Certain jurisdictions, primarily China and the UK, have developed a framework for global FRAND rate determinations, whereby courts identify the rate and other terms that should govern a worldwide SEP license. In other jurisdictions, such as Germany or the UPC, courts do not determine license terms, but assess whether a particular offer, counteroffer, security payment, or course of conduct can be regarded as FRAND (or within a FRAND range, or “not manifestly non-FRAND”).

The second line of jurisprudence treats FRAND primarily as the result of a negotiation process and focuses on the conduct of negotiations. On this view, a FRAND rate and other terms of a license cannot be objectively determined by a court, but can only be established through a negotiation between a willing licensor and a willing licensee in the licensing market.

Despite its inherent conceptual appeal, the implementation of this conduct-based approach has not been universally satisfactory. In our empirical review of the German caselaw on SEP injunctions that we published two years ago (click here to view the paper), we observed that the extent to which German courts (and, in particular, the regional court of Munich) engaged with the substance of licensing offers had significantly declined over time. Instead, courts increasingly focused on behavioral indicators of implementers’ willingness to license, resulting in an almost universal failure of implementers’ FRAND defenses.

In the two years since the publication of our paper, there have been three significant developments. First, the European Commission attempted to influence the application of the Huawei v. ZTE framework through an amicus curiae intervention in VoiceAge EVS v. HMD. In its brief to the Higher Regional Court of Munich, the Commission took the view that the steps of the Huawei v. ZTE framework must be assessed sequentially – and in particular, that the assessment of the substance of the SEP owner’s licensing offer cannot be skipped because of how the implementer responded to that offer. The German Federal Court of Justice, however, confirmed its previous approach in VoiceAge EVS v. HMD, holding that “it is to be assumed without any reasonable doubt that EU law (article 102 of the TFEU) does not lay down any fixed sequence of procedural steps to which strict adherence is required in every case”, and that the FRAND defense must be evaluated through “assessment of the defendant’s overall conduct”.¹⁰ Together with the abandonment of the Commission’s proposed SEP regulation, this outcome suggests that policy interventions have so far had limited success in changing how FRAND disputes are assessed in practice.

Second, the UPC has entered the scene and issued a first set of substantive decisions in FRAND disputes. In particular, the UPC’s Mannheim Local Division charted a new approach in Panasonic v. Oppo. Working within the Huawei v. ZTE framework, the court described that framework as a “negotiation programme” (“Verhandlungsprogramm”) aimed at the timely conclusion of a privately negotiated FRAND license.¹¹ The court emphasized that the parties’ conduct must be assessed in light of the purpose of the negotiation process, namely, whether their behavior contributes to or obstructs the conclusion of a FRAND agreement. Importantly, the court also held that willingness cannot be assessed independently of the substance of the other side’s offer. If there is sufficient initial willingness to license, the SEP holder’s offer must be reviewed for FRAND compliance, and that review must not be omitted and cannot be carried out only cursorily.¹²

Third, and most recently, the Munich Regional Court has moved toward a more valuation-oriented form of review of the negotiation conduct, sometimes combining an assessment of parties’ licensing conduct with an indicative FRAND rate. In Wilus v. ASUSTeK, Broadcom v. Renault, and Nokia v. ASUS, the court examined both the implementer’s willingness and the FRAND-conformity of the SEP holder’s offer.¹³ In doing so, it moved beyond the purely conduct-oriented focus of earlier German caselaw, and also provided relatively specific methodological guidance for its assessment of the substance of licensing offers. In Broadcom v. Renault, the court described a FRAND range extending from 50% below to 50% above a calculated midpoint. In Wilus v. ASUSTeK, it indicated that a top-down cross-check may be available where the SEP holder accounts for a sufficiently large share of the total SEP stack, which, for a large standard, may require at least 1% of the relevant SEPs. This points toward a lighter, faster version of rate analysis; which rests to a considerable extent on rules of thumb.

The most recent Munich decisions in Samsung v. ZTE further illustrate this development.¹⁴ Specifically, the court started from a top-down royalty calculation of USD 1.1 billion for the five-year term from 2024 to 2028. It then applied a 30% volume discount, a 10% increase for Samsung’s alleged holdout, and deducted the value of Samsung’s license to ZTE’s patents. This led to a midpoint value of USD 726 million for a five-year cross-license and an upper end of the FRAND range of USD 798.6 million. ZTE’s offer, therefore, fell below the upper end of the range, so that Samsung could not successfully rely on the FRAND defense in Munich.

The recent developments suggest that the caselaw is moving towards more substantive “FRAND assessments”, where the economic assessment of the parties’ licensing offers is an important element of a court’s assessment of parties’ conduct during the negotiations. From the point of view of FRAND determination methodology, this development raises many new questions. In particular, there seems to be a tension between two different approaches: in the first approach, the assessment of the substance of a party’s licensing offer is contingent on the specific negotiation context. Thus, whether an offer is FRAND depends on the state of the negotiation, the conduct of the other side, and the information that was available to the party at the time it made the offer. In the second approach, the assessment of the FRAND character of the offer ultimately requires the court to conduct an assessment that resembles a FRAND rate determination, as it is on the basis of a specific rate that the court determines specific boundaries of the FRAND range. This stands in contrast to earlier German decisions, in which courts found offers to be within or outside the FRAND range, without necessarily specifying the exact boundaries of that range.

While the assessment of FRAND offers under the Huawei v. ZTE framework uses the same general methodologies that courts in China and the UK have used for FRAND rate determinations, the procedure and the data that are available for the determination remain very different. To substantiate the FRAND character of their licensing offers, parties will need to make the best possible use of the information that is available to them at any given stage of a negotiation, and also during litigation, their access to information will be more limited than in a UK litigation with extensive disclosure of both parties’ relevant information. Thus, the observed trend of courts’ deeper engagement with the substance of negotiating parties’ licensing offers may inspire new ways of using publicly available information, commercial datasets, and tools.

5. Conclusion

The overall picture is one of both convergence and divergence. Courts increasingly agree on the broad meaning and purpose of FRAND, and the focus has shifted to how FRAND should be applied in practice. There is also significant convergence on the main methodologies, especially comparable licenses and top-down analysis. Nevertheless, significant challenges remain at the operational level, including how courts select and unpack comparables, and how they use patent data, e.g., how they account for portfolio strength, essentiality, validity, and geographic coverage. At the same time, there is divergence in the type of review courts undertake: some jurisdictions focus on FRAND rate determinations, while others rely on assessments of parties’ licensing offers in injunction proceedings. Recent developments in Munich and at the UPC suggest that courts are increasingly seeking intermediate forms of review that combine an assessment of parties’ conduct with some level of review of the economic substance of the offers. These new forms of substantive review will call for new approaches to identifying relevant information and data.

About the Study ‘Methodologies for FRAND determination: evidence from global case law’

The new EPO Observatory study, prepared in collaboration with BRELA, reviews 65 FRAND-related court decisions across seven jurisdictions. It provides a structured overview of how courts determine and assess FRAND licensing terms for SEPs, and is intended to support negotiations, litigation, mediation and arbitration. Access the full study.

About the EPO’s ‘Patents and standards programme’

The ‘Patents and standards programme’ at the EPO Observatory on Patents and Technology, aims to improve transparency and predictability in the relationship between patents and standards. The programme pursues dialogue with a broad range of stakeholders, including SEP implementers, SEP holders, licensing pools and standard development organisations. Following the publication of “Standards and the European patent system” and the launch of the Patents Standard Explorer, the present study turns to the licensing of standard-essential patents.

About BRELA

BRELA (Research in Economics and Legal Analytics), provides evidence-based research and expert analysis at the intersection of technology, law, and policy. The firm specializes in the empirical analysis of technology standards, patents, SEP licensing, and FRAND determinations, turning large technical and legal datasets into clear, decision-relevant insights for corporations, law firms, standards organizations, and public-sector institutions, including the European Commission and the European Patent Office.

Beyond data-driven studies and consultancy, BRELA provides independent, methodologically rigorous expert testimony in complex SEP and FRAND matters. BRELA’s Director, Dr. Justus Baron, recently acted as expert witness in the UK FRAND proceedings Samsung v. ZTE, where the Court described his evidence as well-reasoned, carefully presented, and unanswered by any opposing expert.

To discuss how BRELA can support your standards, IP, or policy questions, contact Dr. Justus Baron at justus.baron@brela-research.com.

About FRANDLY

FRANDLY is a licensing intelligence solution for SEPs that combines two complementary offerings. The platform delivers granular SEP licensing data, including royalty benchmarks, market analytics, and supporting documentation, while our research reports provide in-depth economic analysis of key SEP licensing questions, such as ‘The Value of 5G’.

Together, these give professionals and legal experts the evidence basis to make informed decisions across SEP negotiations, litigation, valuation, and policy development. Our data and analyses support reasoned, defensible determinations of FRAND licensing terms.

Access ‘The Value of 5G’

For a comprehensive view of how 5G shapes SEP licensing and product markets, contact us at founders@frandly.com to request the free executive summary or learn more about purchasing the full report. Spanning over 200 pages, it sets out the methodology, descriptive statistics, results, and robustness checks for both the technology and product market analyses.

1. 1. Samsung v. ZTE, case [2026] EWHC 999 (Pat), case number HP-2024-000044, In the High Court of Justice, Business and Property Courts of England and Wales. ↩
   2. ZTE v. Samsung, Chongqing First Intermediate People’s Court. See the Substack of Michael Ma for a link to the document: https://michael7924.substack.com/p/breaking-chongqing-court-ruling-adopts. Alternatively, click here. ↩
   3. LG München I, Endurteil v. 30.04.2026 – case 7 O 64/25. URL: https://www.gesetze-bayern.de/…. ↩
   4. UPC Mannheim Local Division UPC_CFI_189/2025, Order issued on 5 June 2026. ↩
   5. See EPO and BRELA study with the title “Methodologies for FRAND determination: evidence from global case law”, at 10. ↩
   6. WIPO. 2026. FRAND Economics: Valuation Methods in Licensing Standard Essential Patents. URL: https://www.wipo.int/…. ↩
   7. See EPO and BRELA study with the title “Methodologies for FRAND determination: evidence from global case law”, at 12. ↩
   8. See EPO and BRELA study with the title “Methodologies for FRAND determination: evidence from global case law”, at 13. ↩
   9. Microsoft v. Motorola, Case C10-1823JLR, [68]. ↩
   10. BGH, Urteil vom 27. Januar 2026, KZR 10/25 – VoiceAge EVS v. HMD, Pressemitteilung Nr. 21/2026. URL: https://www.bundesgerichtshof.de/…. ↩
   11. Panasonic v. Oppo, Local Division Mannheim, case UPC_CFI_210/2023, [201]. ↩
   12. Panasonic v. Oppo, Local Division Mannheim, case UPC_CFI_210/2023, [198]: “Der Europäischen Kommission ist darin beizutreten …”. ↩
   13. For the three Munich Regional Court decisions, see Wilus v. ASUSTeK, case 7 O 5007/25; Broadcom v. Renault, case 7 O 7655/25; Nokia v. ASUS, case 7 O 4102/25. ↩
   14. See Samsung v. ZTE, Munich Regional Court, cases 7 O 64/25 and 7 O 2750/25. ↩